The IDARESA Data Mediation Architecture for Statistical Aggregates

An approach to meta-data based derivation of statistical aggregates from a “virtual” repository of datasets originating from various pre-existing data sources is presented. This repository provides a browsing and query interface to source systems in terms of a federation platform displaying available information semantically and structurally unified, and a mediation layer making source content look coherent and harmonized at platform level. By strictly formalizing (meta-) information, a synchronized data/meta-data transformation model is built into the query interface. Further, on top of this model, a query resolution algorithm utilizes structural descriptions of statistical aggregates for deriving the requested information from mediated sources. Note: This paper originally appeared in Research in Official Statistics, Vol. 3 (1), pp. 7–38, 2000. However, the printed version deviates from the revised submitted paper and, moreover, contains a whole lot of typesetting errors. This version for download here is the correct one as intended for publication. page 2 M. Denk, K.A. Froeschl: Research in Official Statistics 3 (1), 2000, pp. 7–38